The Pill That Could

Chapter 5: The Long Argument

In Brian Druker's clinic at Oregon Health Sciences University, in the autumn of 1996, a woman is sitting in the chair across from him. She has CML. She was diagnosed eighteen months earlier. She is forty-seven. She has two children in college. She would like to know, given everything Dr. Druker has been able to read in journals and hear at conferences, whether there is something coming.

Druker has, in May of this year, published a paper in a journal called Nature Medicine. It is titled “Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells.” The paper has eight authors. Druker is the first; Nick Lydon, his collaborator at the Swiss pharmaceutical company that supplied the molecule, is the senior author; the others are mostly lab members from the two sides. The paper shows what Druker has been seeing in his lab for two years — a small molecule, called CGP 57148 in the paper, kills BCR-ABL-positive cells and spares normal cells, in cell culture and in mice. The data are clean. The implications are clear.

Druker has, on the bench down the hall, in the freezer, a few hundred milligrams of the molecule.

He cannot give any of it to the patient in the chair. The molecule is owned by the Swiss company. The Swiss company has not yet decided to develop it as a drug. Without that decision, there is no manufacturing at scale, no formal toxicology studies, no regulatory dossier, no clinical trial, and no way for any patient anywhere to take the compound.

Druker does not, in 1996, have a confident answer to give the patient about when this might change.

* * *

It is worth being clear, before going further, about what actually has to happen between a clean result in a dish and a pill in a patient's hand.

There is, first, formal toxicology. A drug candidate has to be tested in animals, generally two species — one rodent (typically rats or mice), one non-rodent (typically dogs, sometimes primates). The animals are given the drug at a range of doses, including doses much higher than what would ever be given to a human, and the people running the studies look at everything they can think to look at: weight, behavior, blood chemistry, organ histology after the animals are sacrificed, the appearance of every tissue under a microscope. The point is to find out what the drug does at high doses to a body, and at what dose problems start to appear. The lower edge of that range — the dose at which the first sign of trouble shows up — sets the upper edge of what regulators will allow you to test in humans, with a built-in safety margin.

There is, second, manufacturing. A clinical trial requires the drug to be made in large enough quantity, with consistent enough purity, that every dose given to every patient is essentially identical. This is harder than it sounds. A molecule that an academic chemist can synthesize a hundred milligrams of, in glassware, with reasonable yield, is not the same problem as a molecule that has to be made by the kilogram, under controlled conditions, with documented batch-to-batch reproducibility. The chemistry has to be optimized, scaled up, run in larger reactors, and validated. The crystalline form of the final molecule — the way the molecules pack together as a solid — has to be selected and locked down, because different crystalline forms of the same compound can dissolve at different rates and produce different drug levels in the bloodstream. The formulation has to be developed: is the drug a tablet, a capsule, a liquid, and what is in there with it? The formulation has to be tested for stability: does the drug still work after it has been sitting in a bottle for six months at room temperature?

There is, third, the regulatory dossier. To start a phase 1 trial in the United States you have to file an Investigational New Drug application — an IND — with the FDA. The IND collects everything: the chemistry, the manufacturing, the animal data, the proposed clinical protocol. It runs to thousands of pages. Reviewers at the FDA read it and either let the trial go forward (typically thirty days after filing, if no concerns are raised) or put it on hold while questions are answered.

There is, fourth, the trial itself. Sites have to be selected. Investigators have to be trained. Institutional review boards — the local ethics committees — have to approve the protocol. Patients have to be identified, enrolled, consented. The drug has to be shipped, stored, dispensed. The data have to be collected and audited. All of this takes money, infrastructure, and people who do this kind of work for a living, which is to say, the resources of a pharmaceutical company.

An academic researcher with a clean result in a dish cannot make any of this happen on their own. They cannot manufacture a clinical-grade kilogram of compound. They cannot file an IND for a molecule they don't own. They cannot, in any practical sense, push a drug forward without the company that holds the molecule deciding to push it forward too.

What an academic researcher with a clean result can do is everything that comes before the company's decision and everything that helps the company change its mind.

All 13 chapters of The Pill That Could · Library · All courses